[REVIEW] Add count_if count_outer_if retrieve_if and retrieve_outer_if functions to static_multiset

[MaybeWilli]

Fixes #800.

Adds if and outer_if functions for count and retrieve functions in static_multiset, to reduce unnecessary hash table lookups when the predicate is false.

Possible future update (left out of this PR for scope):
-Benchmark using

auto const always_true = cuda::constant_iterator<bool>{true};
cuda::std::identity{}

alongside count_if and retrieve_if, similar to how insert_if_n is being used, so that count and retrieve in kernels.cuh can be removed.

[copy-pr-bot]

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

[PointKernel]

@MaybeWilli Thanks for the contribution. There are a few cleanups I need to take care of first before we can move forward with this. Given how our downstream use cases have evolved, we probably no longer need to support the “outer” logic on the cuco side. I’ll let you know once this is ready to move forward.